--- title: "Issue 21" canonical: "https://current.tinyfish.ai/issue/21" publication: "TinyFish" publisher: "TinyFish" language: "en-US" guide: "https://current.tinyfish.ai/llms.txt" permalink: "https://current.tinyfish.ai/issue/21" issue: 21 description: "Agents are inheriting every assumption the web made about humans — the standardized bugs, the undocumented workarounds, the persuasion built for minds that no longer show up." authors: ["Sable Whitford","Rina Takahashi","Nora Kaplan"] published: 2026-03-22 updated: 2026-11-05 articles: 84 --- # TinyFish > CURRENT is a journal about AI agents in production, for the engineering and product leaders who have to make them work. Drawing on experience building enterprise web agents, it explains why the demo succeeded and the deployment didn't, and what that means for choosing vendors, briefing the board, and avoiding mistakes others have already made. [Read on the website]() ## Issue 21 Agents are inheriting every assumption the web made about humans — the standardized bugs, the undocumented workarounds, the persuasion built for minds that no longer show up. ### [Daily Brief]() Your daily digest of web developments through the lens of enterprise agent deployment - [Google Lets Schools Keep Search Without Gemini. The AI Overviews Come Along Anyway.]() — By [Sable Whitford]() · Published 2026-11-04 On Tuesday, which was also Election Day, Google announced a K-12 policy that switches Gemini, Gemini Notebook and **Search off by default** for schools using its education software. A Google education VP told the New York Times that "Search and AI are deeply integrated," so districts that ban AI would lose Search as well. Bernie Sanders called it **"extortion,"** and Sen. Richard Blumenthal and NYC Mayor Zohran Mamdani joined in. Within hours Google **partly walked it back**: schools can now turn on Search by itself, without Gemini. That Search still puts AI Overviews on the results page, though, so **the fight now is over what "AI off" means.** - **Proxy Traffic Leaving the White House Network?** — [Source]() — Published 2026-11-04 Layer3 Intel says residential proxy traffic exited AS2650, the Executive Office of the President's network, 236 times in 30 days through 28 proxy services. It says it wasn't even looking. Its guess is that someone working late has a device carrying a proxy SDK. Treat that as a theory: no device has been named, and six public IPs can sit in front of a lot of machines. - **DreamHost Had Uninvited Guests Since August** — [Source]() — Published 2026-11-04 DreamHost confirmed that attackers used a file manager bug to steal session tokens, customer emails and files, and that they got in back in August. That's a long stay at a host with about 400,000 paying customers. DreamHost says it's unrelated to PHP2Shell. If you're a customer, rotating credentials today is a sensible move. - **PHP2Shell's Hangover Gets a Sequel** — [Source]() — Published 2026-11-04 PHP2Shell was a pair of 9.2-severity bugs that put NASA, Wikipedia and the White House at risk. Two weeks after the patches, Kaspersky counts nearly 30 million servers still unpatched, plus a third, still-open flaw (CVE-2026-115560). Working exploits keep multiplying, though there's no confirmed mass-breach wave yet. Tolmo's Matt Suiche says AI bug hunting is exposing 40 years of legacy code, and that's today's big argument. AI can find a bug in minutes. Someone still has to book maintenance windows for 30 million servers. - **Can DNS Tell You It's a Bot?** — [Source]() — Published 2026-11-04 Singapore's IMDA is reportedly exploring, with GoDaddy, Cloudflare and Google, whether DNS could identify AI agents. That would give the internet's 1983 phone book a second job checking IDs at the door. Every ops person already knows where the postmortem will land: it's always DNS. - **Meta Buys Conductor, Cloud Users Get Notice** — [Source]() — Published 2026-11-04 Meta bought Conductor, the app for running several coding agents in parallel, and the team is joining Meta Superintelligence Labs. Terms weren't disclosed. Conductor Cloud goes read-only Nov. 13 and shuts off Nov. 20. The local app gets model updates and critical fixes until early 2027. Conductor ran agents from rival labs too, so everyone wants to know whether that survives the move. - **OpenAI Shops Too: Pierre Computer** — [Source]() — Published 2026-11-04 Not to be outdone, OpenAI told Axios it plans to buy Pierre Computer, which stores and manages AI-generated code. It's OpenAI's second coding pickup after the Ona deal announced in June. Agents are producing mountains of code nobody typed by hand, and someone has to keep track of it. OpenAI would like to be that someone. - **Kimi K3.1 Pro Weights Drop: Bring Disks** — [Source]() — Published 2026-11-04 Moonshot launched Kimi K3.1 Pro in October, and today the open weights landed on Hugging Face. It has 2.8 trillion parameters with 104B active, and reportedly ships as 96 shards totaling about 1.56 TB under a custom non-OSI licence. Self-hosting starts with a purchase order for storage. Before the release, Artificial Analysis said it would cut the gap between open and proprietary models from 12 points to 3. - **Ripple-65B Crowns Itself America's Coding Champ** — [Source]() — Published 2026-11-04 Ricursive calls its Ripple-65B Preview "U.S. Open-Weight SOTA on Agentic Coding," meaning best in its class. That's the company grading its own homework. The "U.S." qualifier is carrying extra weight today, since a Chinese lab just released 1.5 terabytes of open weights. - **Go Gets a Brand-New JSON API** — [Source]() — Published 2026-11-04 The Go team introduced encoding/json v2, a new official JSON API, and Gophers have plenty to say about it. Nearly every Go service reads or writes JSON somewhere, so a second version of the standard package lands on the desk of anyone shipping APIs. - **Apple Puts Engineers on an AI Diet** — [Source]() — Published 2026-11-04 Bloomberg reports that in October, Apple cut some engineers' daily AI spend caps from $300 to $200. Some could drop to $100, and certain non-engineering teams could be capped at $25. Some teams are being nudged toward cheaper or on-device models. The punchline: few people ever hit $300, and Apple's R&D rose 32% in fiscal 2026. - **Texas Voted, and Data Centers Flinched** — [Source]() — Published 2026-11-04 Last brief, the data center backlash was winning local ballot fights. Now it has the Texas governor's mansion, in a state holding 25%+ of planned US capacity. Gina Hinojosa ran on an immediate moratorium, tougher water and power rules, and no tax breaks, and AI data center stocks fell on the result. GLJ Research says about 13 GW is exposed. Altimeter's Brad Gerstner says 10–15 GW might never get built. Abbott had already paused permits. Musk, who spent big on the losing side, now likes "data centers in space." - **Big Tech Pulls Out the AI Credit Card** — [Source]() — Published 2026-11-04 Oracle filed to sell $20B of new stock and bonds for its AI buildout, and Meta is seeking about $25B in bonds. The timing is bold. They're raising money to build on the same day the state with the most planned capacity elected a governor who ran on a moratorium. - **ChatGPT Search Is 20% of Google's?** — [Source]() — Published 2026-11-04 OpenAI's Adam Fry said Monday that ChatGPT search volume is about 20% of Google's, and that weekly users reportedly went from 1B to 1.2B in 60 days. Check the counting rules first. Any prompt that triggers a web search counts, and OpenAI didn't disclose the Google baseline or its method. Ahrefs puts ChatGPT at 0.36% of web referral traffic. - [Faria Huq and the Reason Behind a Human Step]() — Published 2026-10-30 When a person takes control from an AI agent driving their web browser, the system records a human step, but not why. Faria Huq and her Carnegie Mellon collaborators have built tools around these moments: an agent that can't find the right forum, a shopper who realized the brand mattered only after seeing the agent's pick. Their system learns to predict when someone will step in. The reason it sets aside would tell a team whether to fix the agent, the website, or its understanding of the person. - [Lebanon's Two Clocks and the Question Calendars Don't Ask]() — Published 2026-10-30 On March 26, 2023, Lebanon woke up on two clocks. The caretaker government had postponed daylight saving three days earlier. Churches and TV channels moved forward anyway, government offices stayed put, and most phones jumped on their own. Middle East Airlines printed every departure an hour earlier, so its planes still left at the same moment. That week made each calendar show what it had been storing: an exact instant, or a time on the wall that politics can move. Software assistants now make that choice without asking anyone. - [Who Pays While the Agent Waits]() — Published 2026-10-30 Google's new service for hosting AI agents starts from a fact anyone who has watched one run already knew: most of the time, the agent is waiting, for the model to answer or for a person to reply. So Google suspends idle agents and promises you won't pay for idle machines. Fair enough, but the wait still gets paid for. It moves into approval queues nobody times, and then onto the plan, which comes back exactly as it left, even if the stock it counted on has since sold. - [Your Agent's Questions Need a Recipient and a Timer]() — Published 2026-10-30 When an AI agent stops mid-task to ask a human something, it often reads as a failure of autonomy. In one study, letting coding agents ask about bug reports stripped of key details won back most of the lost ground, lifting solve rates from about 55% to 69.4%. But the "human" answering was another model, holding the full report and standing by. Real people are in meetings, or asleep. The design work is in the ask: who gets it, what it says, and what happens while it waits. ### [Market Pulse]() Reading the agent ecosystem through a practitioner's lens - [Every Layer Works. Nobody Owns the Outcome.]() — By [Rina Takahashi]() · Published 2026-03-19 · Format: The Stratification Problem At GTC last week, NVIDIA unveiled kernel-level sandboxing for agent runtimes. In the same window, Google and Microsoft pushed WebMCP into W3C incubation. Visa and Mastercard each advanced competing protocols for agent-initiated payments. Three layers of the agent stack, three governance models, each technically serious. Each also scoped to exactly what its owner can control. When an agent researches, interacts, and transacts across all three layers, who's accountable for the outcome? The dashboards all look green. - **NVIDIA Unveils Full Agent Runtime Stack at GTC** — [Source]() — Published 2026-03-19 GTC brought the OpenShell runtime, CPU-optimized Vera Rubin racks for agent workloads, and an Agent Toolkit with 15+ enterprise partners including SAP, Salesforce, and ServiceNow. The NemoClaw/OpenClaw security framing positions NVIDIA as the sandbox layer for enterprise agents, though NemoClaw remains alpha-stage. - **WebMCP Now Testable in Chrome Canary Build** — [Source]() — Published 2026-03-19 The WebMCP protocol, co-developed by Google and Microsoft, is live behind a flag in Chrome Canary. It lets agents interact with websites through declared structured APIs instead of pixel-scraping. The spec is incubating through a W3C community group, with formal browser announcements expected mid-to-late 2026. - **Visa and Mastercard Ship Competing Agent Protocols** — [Source]() — Published 2026-03-19 Both networks launched agentic commerce frameworks in the same quarter, each building on Cloudflare's Web Bot Auth cryptographic layer. Visa is coordinating with 100+ partners and aligning with OpenAI and Coinbase standards. Mastercard's Agent Suite arrives Q2. Liability allocation remains conspicuously unaddressed. - **Context Rot Research Punctures Long-Context Window Claims** — [Source]() — Published 2026-03-19 Research from Chroma and others finds models advertising 1M+ token context windows suffer severe performance degradation well before their limits, with effective capacity typically topping out around 60–70% of the headline number. Semantically similar distractors make things worse, which is bad news for agentic coding tasks. - **Claude Opus 4.6 Ships With Million-Token Window** — [Source]() — Published 2026-03-19 Anthropic's latest Opus release brings a generally available 1M token context window and leads Terminal-Bench 2.0 for agentic coding at 65.4%. A new context compaction feature automatically summarizes older context to sustain longer sessions. Parallel subagent teams arrive as a research preview. - **Luma Agents Orchestrate Multi-Model Creative Production Pipelines** — [Source]() — Published 2026-03-19 Luma launched Agents, a platform coordinating 8+ models across text, image, video, and audio from a single brief. Publicis Groupe, Adidas, and Mazda are already running on it. The CEO's claim that one campaign dropped from $15M to under $20K remains unaudited. - [Visa and Mastercard Ship Competing Agent Payment Protocols While the Liability Framework Remains Unwritten]() — Published 2026-03-19 Both card networks shipped agentic commerce frameworks in the same quarter. Visa's Intelligent Commerce spans over 100 partners. Mastercard's Agent Pay completed Australia's first authenticated agent transaction in January. The protocol infrastructure is moving. - **Context Rot: How Increasing Input Tokens Impacts LLM Performance** — [Source]() — Published 2026-03-19 Across 18 frontier models, Chroma's research team found universal performance degradation well before advertised context limits. Effective capacity runs roughly 60–70% of maximum. Counterintuitively, models handled shuffled text better than coherent text, implicating attention architecture itself. - **Improving LLM Function-Calling via Guided-Structured Reasoning Templates** — [Source]() — Published 2026-03-19 Proposes structured reasoning templates that walk models through step-by-step function-call generation, tackling documented 20–27% accuracy drops caused by rigid JSON output constraints. Achieves 3–12% relative improvements over strong baselines while making tool-use decisions more interpretable. - **Multi-Agent Deep Reinforcement Learning Under Constrained Communications (DG-MAPPO)** — [Source]() — Published 2026-03-19 A fully distributed multi-agent RL framework that strips out centralized training dependencies via a Distributed Graph Attention Network. Agents learn to coordinate through peer-to-peer communication alone, outperforming centralized baselines on StarCraft II and Google Research Football. - **Agentics 2.0: Logical Transduction Algebra for Agentic Data Workflows** — [Source]() — Published 2026-03-19 March 2026 preprint formalizing each LLM call as a typed semantic transformation composable through algebraic operators. These execute as stateless asynchronous calls in parallel Map-Reduce architectures, opening a path toward formal verification of agent behavior. - **Deeptune Raises $43M Series A for High-Fidelity RL Training Environments** — [Source]() — Published 2026-03-19 Deeptune closed a $43M Series A led by a16z to build reinforcement learning environments simulating professional workflows at high fidelity. The investment thesis: the bottleneck for capable agents is training infrastructure, not model architecture alone. - **Nvidia, the world's most valuable company, just sent another signal that AI agents are going to be everywhere** — [Source]() — Published 2026-03-19 CNN reads GTC's hardware bets as confirmation that agent-specific compute is hardening into its own infrastructure layer. - **The Race to Bank AI Agents in New Era of Commerce Is Heating Up** — [Source]() — Published 2026-03-19 Bloomberg tracks how programmable money for agents turned the commerce layer into contested financial terrain overnight. - **WebMCP turns any Chrome web page into an MCP server for AI agents** — [Source]() — Published 2026-03-19 The New Stack surfaces a design nuance most coverage missed: WebMCP's protocol layer assumes co-piloting, not autonomy. - **NVIDIA Expands Enterprise AI Push with OpenShell and Agent Software** — [Source]() — Published 2026-03-19 ADTmag details the runtime governance layer beneath GTC's headlines, plus the analyst caveat that it isn't complete. - **Mastercard builds a payments AI model** — [Source]() — Published 2026-03-19 Mastercard's large tabular model signals that payments incumbents are building proprietary infrastructure at the commerce layer. - **Google Chrome ships WebMCP in early preview, turning every website into a structured tool for AI agents** — [Source]() — Published 2026-03-19 VentureBeat's deep dive remains the clearest map of how WebMCP reframes the protocol layer from tolerating scraping to declaring intent. ### [Foundations]() Conceptual clarity earned from building at scale - [Where Intelligence Earns Its Keep in Browser Automation]() — By [Nora Kaplan]() · Published 2026-03-19 · Format: Hybrid By Design A browser automation workflow runs for the hundredth time. One open-source framework barely touches a language model. The other sends the full page state to one at every step, same as it did on run number one. Same task, same browser. Wildly different bills at the end of the month. - [The Flight Simulator Problem]() — By [Rina Takahashi]() · Published 2026-03-19 · Format: Learning By Doing More than 35 companies are now building practice environments for AI agents. Anthropic has discussed spending over a billion dollars on them. The pitch is intuitive: you wouldn't let a pilot fly without simulator time. Deeptune just raised $43 million from a16z on exactly this logic, building pixel-perfect replicas of Slack, Salesforce, and ticketing tools where models can rehearse. But flight simulators work because physics holds still. The live web does not. And there's a subtler problem inside these training environments, one that lives deeper than fidelity. - **The 2026 MCP Roadmap** — [Source]() — Published 2026-03-19 The protocol maintainers name the production friction points they haven't solved yet. That candor is rare. - **Andreessen Horowitz Backs Deeptune's $43M Bet on Training Gyms for AI Agents** — [Source]() — Published 2026-03-19 The Fortune exclusive that put RL simulation environments on the map as a standalone infrastructure category. - **NVIDIA Launches Enterprise AI Agent Platform at GTC 2026** — [Source]() — Published 2026-03-19 Seventeen partners, one platform thesis, and carefully hedged language worth reading between the lines of. - **Migrate Your Semantic Kernel and AutoGen Projects to Microsoft Agent Framework** — [Source]() — Published 2026-03-19 The API surface is frozen and AutoGen's future is maintenance-only. If you're facing this transition, start here. - **MCP vs A2A: Understanding How the Two Protocols Complement Each Other** — [Source]() — Published 2026-03-19 One protocol is vertical, the other horizontal. This is the clearest explanation of why they don't actually compete. - **The State of AI and Browser Automation in 2026** — [Source]() — Published 2026-03-19 A vendor landscape view of how browser automation infrastructure is consolidating. Read for the framing, not the statistics. - **Stagehand V3 Drops Playwright for Raw CDP** — [Source]() — Published 2026-03-19 V3's big architectural bet: swapping Playwright internals for direct Chrome DevTools Protocol access, which buys a 44% speed gain on gnarly DOM structures like shadow roots and iframes. Element caching means repeatable actions skip the LLM entirely. AI only fires when the script can't cope. - **Browser Use Lets the Model Drive Everything** — [Source]() — Published 2026-03-19 At 78,000+ GitHub stars, Browser Use takes the opposite position: hand an LLM a goal and let it reason its way through. Every single step involves live inference, blending DOM parsing with screenshot analysis. No pre-programmed paths. It adapts on the fly, trading predictability for autonomy. - **Playwright Remains What Everything Else References** — [Source]() — Published 2026-03-19 No AI, no inference calls. Playwright controls Chromium, Firefox, and WebKit through explicit selectors and scripted actions. With 78,600+ stars and 45.1% QA adoption, it's the deterministic bedrock. Now AI-connectable through MCP, it's quietly becoming the stable layer agent frameworks orbit around. - **Teams Keep Landing on an Eighty-Twenty Split** — [Source]() — Published 2026-03-19 The pattern production teams keep discovering independently: Playwright handles the roughly 80% of steps that are predictable, AI frameworks cover the 20% that require genuine understanding. Pure AI wobbles at scale. Pure scripting shatters when UIs shift. The hybrid works because both failure modes are real. - **Deeptune Raises $43M for Agent Training Gyms** — [Source]() — Published 2026-03-19 Andreessen Horowitz leads Deeptune's Series A to build reinforcement learning environments that simulate enterprise workflows across tools like Slack and Salesforce. The signal: RL training is becoming its own infrastructure category, bridging the gap between raw model capability and agents that can actually do specific jobs. - **Kai Exits Stealth With $125M Security Platform** — [Source]() — Published 2026-03-19 Evolution Equity Partners leads Kai's combined seed and Series A for a cybersecurity platform built from scratch around agentic threats. The signal: security and governance tooling has graduated from "nice to have" to primary investment thesis, with AI-driven attacks demanding AI-native defense. - **Axiom's $200M Bet on Verifiable AI Code** — [Source]() — Published 2026-03-19 Menlo Ventures leads Axiom's Series A at a $1.6B valuation for systems that mathematically verify code using Lean proof language. The signal: as AI writes more production code, formal verification is shifting from optional quality assurance to compliance prerequisite. Trust, made auditable. - **Oro Labs Lands $100M for Agentic Procurement** — [Source]() — Published 2026-03-19 Brighton Park Capital and Goldman Sachs Growth Equity co-lead Oro Labs' Series C after 300% revenue growth. The signal: vertical agentic apps with measurable, workflow-specific ROI are proving out as the enterprise adoption path investors trust most. Goldman's involvement says something about who's paying attention. - **Gartner Projects 40% Agent Penetration by 2026** — [Source]() — Published 2026-03-19 Gartner predicts 40% of enterprise applications will feature task-specific AI agents by end of 2026, up from under 5% in 2025. The signal: an eightfold jump in a single year. This week's infrastructure bets are positioning ahead of a curve that's already bending sharply upward. ### [Practitioner's Corner]() Lessons from the field—what we see building at scale - [The Next Translation Layer]() — By [Nora Kaplan]() · Published 2026-03-19 · Format: The Translated Dependency The IRS's Individual Master File has been running since 1961. It was supposed to be replaced decades ago. Instead, layer after layer of new technology grew around it—databases, portals, APIs—each one promising to make the last one obsolete. None did. And at some point, the logic that kept things working migrated out of the code and into the people operating it. The workarounds became the process. - [Building Browser Agents for a Web That Fights Back]() — By [Rina Takahashi]() · Published 2026-03-19 · Format: Building for Breakage A government portal in Delaware goes offline at night. A dropdown is actually a textbox. A checkbox arrives pre-checked, hoping nobody notices. The web is a collection of pages to read and forms to fill. It is also, for anything trying to automate those forms, an adversary: elements misidentified, structures obfuscated, layouts rearranging between visits. Most browser agents get built as though the environment cooperates. - [I Am the Diagram]() — Published 2026-03-19 *The following interview is a composite construction. Marisol Vega-Lundquist is not a real person. She is built from publicly documented patterns of government IT modernization failure, GAO reports, and the operational realities described by actual state workforce agency officials during and after the COVID-19 unemployment surge. Her specifics are invented. Her situation is not.* - [Amazon's AGI Lab Is Building Faithful Copies of Broken Software to Train Its Most Advanced Agents]() — Published 2026-03-19 Every enterprise AI pitch starts the same way: modernize the stack, add APIs, escape the legacy mess. Amazon's AGI Lab is running the opposite play. Its researchers build reinforcement learning gyms that faithfully reproduce decades-old software, quirks included. Modal windows appearing late. Fields rejecting input until some other value saves first. Forms silently resetting midflow. - **How Agentic AI Helps Heal the Systems We Can't Replace** — [Source]() — Published 2026-03-19 Reframes legacy system quirks as training signal, not obstacles. The sharpest inversion of the replacement narrative this week. - **5 AI Agent Failure Patterns and Production Fixes** — [Source]() — Published 2026-03-19 Practitioner taxonomy of what breaks when agents meet production logic they weren't trained to expect. - **Agentic AI Is Forcing Analytics and Operations to Converge** — [Source]() — Published 2026-03-19 Agents dissolve the workload boundaries enterprise data architecture was built around. The infrastructure consequences are just starting. - **The State of AI and Browser Automation in 2026** — [Source]() — Published 2026-03-19 Governance and observability layers between browser agents and legacy portals remain mostly human-shaped. Useful ecosystem snapshot. - **Why AI Pilots Fail in Production and the 2026 Integration Roadmap** — [Source]() — Published 2026-03-19 Frames the gap as an OS problem: models work, but the operating layer for enterprise integration doesn't exist yet. ### [Echoes]() Past infrastructure decisions echoing in today's production realities - [The Specification of Being Wrong]() — By [Nora Kaplan]() · Published 2026-03-19 · Format: The Standardized Bug In 1996, the W3C published a specification telling browsers they were calculating the width of a box wrong. The browsers had shipped first. Millions of pages already depended on the "incorrect" math. Somebody had to give. - [The Typo Every Browser Still Sends]() — Published 2026-03-19 In 1995, someone on an IETF mailing list pointed out that the HTTP "Referer" header was misspelled. Phillip Hallam-Baker, who'd proposed the field, cracked a joke about Daleks. Roy Fielding noted the Unix spell checker of the era wouldn't have caught either spelling. Everyone shrugged. - [The Woman Who Keeps the Web Wrong]() — Published 2026-03-19 Every major browser engine shipping today contains a rendering path whose explicit purpose is to be wrong. Wrong in the specific ways that Netscape Navigator 4 and Internet Explorer 5 were wrong, preserved with the fidelity of a museum restoration. - [The Fifty-Year Byte]() — By [Rina Takahashi]() · Published 2026-03-19 · Format: Permanent Decisions In the early 1970s, someone at Bell Labs made a quiet decision about how C would store text. No memo survives. That choice created the structural conditions for buffer overflow vulnerabilities, a category of harm that was understood by 1988, studied exhaustively through the 1990s, and the target of a billion-dollar defensive industry by the 2000s. None of it was enough. The flaw persisted for fifty years. Understanding arrived in the first twenty. - [The Clock Everyone Can See]() — By [Rina Takahashi]() · Published 2026-03-19 · Format: Permanent Decisions At 03:14:07 UTC on January 19, 2038, a counter running since 1970 will hit its maximum value and wrap to a date in December 1901. The math is simple. The fix exists. Everyone who needs to know, knows. Y2K mobilized hundreds of billions in global remediation through deadline pressure and concentrated blame. The 2038 problem has the same clarity and none of the same urgency. Twelve years out, the gap between knowing and doing has barely narrowed. - **Back to the Building Blocks: The White House ONCD Technical Report on Memory-Safe Languages** — [Source]() — Published 2026-03-19 The federal government names C's memory model as a structural debt class spanning thirty-five years of vulnerabilities. - **DARPA's TRACTOR Program: Automated C-to-Rust Translation as a National Security Project** — [Source]() — Published 2026-03-19 When a language's legacy becomes a defense liability, DARPA funds machine translation at infrastructure scale. - **Memory-Safe Migration Will Take a Generation, Not a Policy Cycle** — [Source]() — Published 2026-03-19 Vaughan-Nichols frames the move away from C as a decades-long replacement problem with no clean cutover. - **The 2038 Problem Already Has a CVE, and Embedded Systems Aren't Patched** — [Source]() — Published 2026-03-19 Kaspersky identifies live vulnerabilities in fuel-gauge consoles as early evidence of 32-bit time overflow arriving. - **The 2038 Bug: Which Sectors Actually Face Risk from Legacy Time Representations** — [Source]() — Published 2026-03-19 Electropages surveys where 32-bit Unix timestamps still sit in production, with useful specificity about embedded systems. ### [Vision]() Where human-AI collaboration is heading - [What Breaks When the Buyer Has No Brain to Hack]() — By [Nora Kaplan]() · Published 2026-03-19 · Format: The Persuasion Stack Ninety-seven percent of popular websites and apps deploy dark patterns—design choices engineered for a distractible, loss-averse, emotionally available human buyer. That might be the commercial web's actual load-bearing structure. Gartner predicts 90% of B2B purchasing will be agent-intermediated by 2028. Early research on how AI agents actually navigate e-commerce is surfacing results that nobody in the $258 billion digital advertising industry is going to love. The interesting part is what's left standing once the persuasion layer falls away. - [When the Customer Has No Eyes]() — By [Rina Takahashi]() · Published 2026-03-19 · Format: New Rules, New Gatekeepers The commercial web was built to persuade people looking at screens. Landing pages, hero images, the choreography of conversion funnels. But a growing share of product discovery now runs through AI agents that don't scan or feel. They parse. Roughly 69% of websites have no structured data for them to parse at all. Being found was always the first problem of commerce. What happens when the thing doing the finding can't see you? - [Who Decides What an Agent Can Trust?]() — By [Rina Takahashi]() · Published 2026-03-19 · Format: New Rules, New Gatekeepers Visibility is the first problem. Trust is the harder one. Once an agent finds a vendor, it needs to decide whether to transact, and the signals humans rely on — brand recognition, a colleague's offhand recommendation, gut feel — don't compute. Someone has to build the infrastructure that tells agents what's trustworthy. History suggests whoever does gains power that outlasts any single transaction. That layer is being poured now, and concrete sets fast. - [A VP of Marketing Watches Her Craft Disappear Into the Void Where Page Two Used to Be]() — Published 2026-03-19 *Margot Vellum is not a real person, though she is, regrettably, a composite of several real conversations that all ended with someone staring into a glass of wine and saying some version of "I don't know what I'm for anymore." She is the VP of Marketing at a mid-size cybersecurity SaaS company, a former tech journalist who crossed into content marketing fifteen years ago and built a career on the conviction that good writing wins deals. We spoke over video. She had a mug that read "Content Is King" with "King" crossed out in Sharpie and replaced with nothing.* - [The First Legal Line Around Agent Commerce]() — Published 2026-03-19 Judge Chesney's March 10 injunction against Perplexity's Comet agent landed on a distinction that will define the next decade of commerce: user permission is not platform authorization. Your agent can act for you. The platform still gets a veto on whether it walks through the door. - **AI Agents as Enablers of Personalized Law: Towards Agentic Disclosures?** — [Source]() — Published 2026-03-19 Oxford legal scholars ask what disclosure law even means when the "reader" is a procurement agent. - **Agentic Commerce Shifts B2B Marketplaces From Intermediaries to Infrastructure** — [Source]() — Published 2026-03-19 PYMNTS on why structured payment terms and clean data now win deals that relationships used to. - **How Brands Are Repositioning for Agentic Engine Optimization** — [Source]() — Published 2026-03-19 The WEF's primer on machine-readable brand narratives, and why search rankings are becoming a trailing indicator. - **How AI Agents Discover B2B Vendors** — [Source]() — Published 2026-03-19 Practical look at agent shortlisting mechanics in procurement. If you're not in the first pass, you're nowhere. - **The Robots.txt for AI Agents Is Coming: What the Open Banking Analogy Reveals** — [Source]() — Published 2026-03-19 A practitioner's case for structured agent access protocols, borrowing hard-won lessons from open banking. - **From Robots.txt to AI Regulation: Web Standards Evolving for Machine Consumers** — [Source]() — Published 2026-03-19 Useful synthesis of IETF, W3C, and EU efforts to rebuild web governance around non-human traffic patterns. ## More - [Publication guide]() - [Issue archive]()